基于图像分析的热电池缺陷无损检测方法

Non-destructive testing method for thermal battery defects based on image analysis

针对目前国内外热电池装配缺陷检测多靠人工分辨的问题,提出了一种基于图像分析检测热电池装配缺陷的方法。通过对热电池的标准单体电池和缺陷单体电池进行分析,以热电池内部结构特征自对比实现对热电池内单体电池的整体倒装、单体电池漏装集流片、装配次序错误进行检测,建立了一个无损检测系统。利用模板匹配将X射线仪拍摄的电池图片中的电池堆部分进行分割,对电池堆不同位置进行竖向扫描产生灰度与位置的波形图,通过自身的对比以及多位置的结构特征参数进行缺陷判断。最终对无损电池检测正确率为94%,缺陷电池检出准确率为96%。

Aiming at the problem of manual battery thermal cell assembly defect detection relying on manual resolution, a method for detecting thermal battery assembly defects based on image analysis is proposed.By analyzing on standard single cell and the defective single cell of the thermal battery, the internal flip-chip of the single cell in the thermal battery, the leakage of the single cell and the assembly order are performed by the internal structural characteristics of the thermal battery, a non-destructive testing system is established.The principle is using template matching to extract battery stack in battery picture taken by X-ray apparatus, and perform vertical scanning on different positions of battery stack to generate waveform diagram of gray scale and position, and perform defects judgement by self-contrast and structural feature parameters of multiple positions.In the end, the correct rate for nondestructive battery testing is 94 %,and the accuracy of defective batteries is 96 %.